The analog formulation of sparsity implies infinite divisibility and rules out Bernoulli-Gaussian priors.
Arash AminiUlugbek KamilovMichael UnserPublished in: ITW (2012)
Keyphrases
- association rules
- rule sets
- sparse approximation
- high dimensional
- maximum likelihood
- classification rules
- real valued
- production rules
- neural network
- mixture model
- gaussian mixture model
- efficient optimization
- gaussian noise
- gaussian distribution
- closed form
- decision rules
- signal processing
- prior knowledge
- bayesian networks